To answer RQ2, experiments have been carried out to identify the prototype performance in laboratory and full-scale flight conditions, as presented in Paper IV. Throughout the project, the total runtime for the two fuel cells is 22 and 24 hours.
3.2.1 Polarization Curves
Fig. 8: Polarization curve showing the current-voltage characteristics (i-V) for the two Aerostak fuel cells.
The highest overall performance measured throughout testing was 27.2 A and 25.9 A at 1176 W and 1133 W for FC A and FC B, respectively.
By running a polarization-curve test and plotting the current I and voltage V values, a polarization plot for the two Aerostak fuel cells was obtained (Fig. 8). This serves as a practical reference for the nominal performance. A simple linear expression (R2 = 0.95) for the fuel cell voltage 𝑉𝐹𝐶 as a function of current 𝐼𝐹𝐶 is given in Eq. 7.
𝑉𝐹𝐶 = 56.445 − 0.5386 ∙ 𝐼𝐹𝐶 (7)
Results
In this test, the fuel cells top out at about 25 A and a voltage of 43 V, giving an average cell voltage of 0.66 V. The power supply voltage was set to 43 V during testing, and this forms a fuel cell output limit and defines when the secondary power source steps in to supply further power. By multiplying voltage and current, the electric power output is obtained. A curve for the power output using the linear expression is included in Fig. 8.
3.2.2 Load Cycle Testing
Fig. 9: Test data from a load cycle with both fuel cells and a constant power supply voltage of 45.1 V.
When exposed to a load cycle with a 2.8 kW take-off phase (Fig. 9), the fuel cells jumped to provide a combined output of 1565 W, which is 78%
of the rated nominal performance. At 30 seconds after take-off, the fuel cells reached 90% of nominal output. The output further climbed towards full power throughout the cruise phase.
The secondary power source served its purpose and provided a power buffer at take-off as the fuel cells ramped up. The peak hybrid power was 1351 W and a current of 32.5 A. For a 16 Ah battery, that would give a peak discharge rate of 2 C. Spikes in power supply contribution
Results
between stack A and B purging is consistent throughout the test. The fuel cells reported 261 Wh of energy, making the secondary power source energy contribution 14 Wh, which is 5% of the load profile total energy.
In Fig. 10, the total fuel cell power at six different power supply voltage levels are presented. This demonstrates how the voltage of the secondary power source, representing different battery state-of-charge levels, influences the fuel cell power contribution throughout the load profile.
This is a key concept utilized in passive hybrid management systems, and from Eq. 7 it can be found that the fuel cell output will vary by 25%
as the hybrid battery state-of-charge is reduced by 3.5 V.
Fig. 10: Combined fuel cell power for a load profile at different power supply voltages. The different voltage levels represent different state-of-charge for 11-cell and 12-11-cell Li-Ion batteries.
At the highest voltage level, 50.4 V, the total fuel cell power is limited to 1200 W and an individual fuel cell contribution of 600 W. The total energy provided throughout the load profile is then 64% of the complete load cycle. When the voltage is lowered to 45.1 V, the fuel cell provides 95% of the energy. Thus, with a passive hybrid strategy, the fuel cell contribution is somewhat limited when the battery state-of-charge is 100% and will increase as the battery discharge. The fuel cell dynamic
Results
response is better when the voltage is high, and the fuel cell loading is lower.
3.2.3 Test Flight Performance
The full-scale test flight phases were: standby, conditioning, take-off, hover, temporary landing with spinning propellers, free flight, and landing. The fuel cells' performance from the flight is plotted in Fig. 11.
The maximum power reported by FC A and FC B was 995 W and 963 W at 24.3 A and 23.2A, respectively. Water drops were found in both fuel cell purge tubes after the flight, indicating adequate hydration levels at landing.
Fig. 11: Fuel cell performance from the test flight with the fuel cell-powered Staaker BG200.
In standby, the relative contribution of FC B drops to zero while FC A takes over and provides all power. As the propellers are started in the conditioning phase, the contribution of FC B increases, but it is not until after take-off that FC B accelerates its power contribution and reaches full output after 5-6 minutes. After the temporary landing, both fuel cells have equal response to the dynamic load at take-off and immediately reach full power. FC A has nominal performance throughout the flight.
Results
The cause of the uneven performance is that FC A had better initial performance and higher voltage. This trapped FC B in a negative loop where FC A continued to improve hydration and performance, further increasing its load share and making it difficult for FC B to catch up.
While there are mitigation strategies, this demonstrates a challenge with system architectures using multiple fuel cells. Individual fuel cell performance is highly dependent on membrane status and will vary as they degrade. A consequence of uneven load distribution is a higher use of the secondary power source, which can lower the energy margins.
During the second half of the hovering phase, it appears that the purging sequence is synchronized between the fuel cells. Since there are slight variations in the purging sequence at low and high power outputs, a purge synchronization can occur when the fuel cells operate at different power outputs. This is unfortunate because the hybrid battery discharge loading doubles when it has to compensate for both fuel cells, increasing the discharge peak currents from 25 A to 50 A. This may impact the overall battery capacity, power stability and flight behavior.